Unfortunately that’s not really relevant to LLMs beyond inserting things into the text you feed them. For every single word they predict, they make a pass through the multi-gigabyte weights. Its largely memory bound, and not integrated with any kind of sane external memory algorithm.
There are some techniques that muddy this a bit, like MoE and dynamic lora loading, but the principle is the same.
A1111
Eh, this is a problem because the “engine” is messy and unoptimized. You could at least try to switch to the “reforged” version, which might preserve extension compatibility and let you run features like torch.compile.
Oh you should be able to batch the heck out of that on a 4080. Are you not using HF diffusers or something?
I’d check out stable-fast if you haven’t already:
https://github.com/chengzeyi/stable-fast
VoltaML is also old at this point, but it has really fast AITemplate implementation for SD 1.5: https://github.com/VoltaML/voltaML-fast-stable-diffusion
Oh, 16GB should be plenty for SDXL.
For flux, I actually use a script that quantizes it down to 8 bit (not FP8, but true quantization with huggingface quanto), but I would also highly recommend checking this project out. It should fit everything in vram and be dramatically faster: https://github.com/mit-han-lab/nunchaku
You don’t want it to anyway, as “automatic” spillover with an LLM painfully slow.
The RAM/VRAM split is manually configurable in llama.cpp, but if you have at least 10GB VRAM, generally you want to keep the whole model within that.
Try a new quantization as well! Like an IQ4-M depending on the size of your GPU, or even better, an 4.5bpw exl2 with Q6 cache if you can manage to set up TabbyAPI.
Depends which 14B. Arcee’s 14B SuperNova Medius model (which is a Qwen 2.5 with some training distilled from larger models) is really incrtedible, but old Llama 2-based 13B models are awful.
But the original Jurassic Park was fun. The writing was sharp and memorable, the cast charismatic, even if the plot is not that important (which is fine).
I did love JC Avatar’s alien flora and fauna, and some small details like the realistic spaceship, but I guess it feels much less exciting in hindsight without anything to “attach” it to.
And again… the IP its name collides with is nothing to sneeze at, visually. You can pause it almost anywhere, even in “mundane” scenes, and get gorgeous fantasy shots and incredible music:
No, all the weights, all the “data” essentially has to be in RAM. If you “talk to” a LLM on your GPU, it is not making any calls to the internet, but making a pass through all the weights every time a word is generated.
There are system to augment the prompt with external data (RAG is one word for this), but fundamentally the system is closed.
There is no movie in Ba Sing Se.
Here, we are safe. Here, we are free.
Oh I didn’t mean “should cost $4000” just “would cost $4000”
Ah, yeah. Absolutely. The situation sucks though.
I wish that the vram on video cards was modular, there’s so much ewaste generated by these bottlenecks.
Not possible, the speeds are so high that GDDR physically has to be soldered. Future CPUs will be that way too, unfortunately. SO-DIMMs have already topped out at 5600, with tons of wasted power/voltage, and I believe desktop DIMMs are bumping against their limits too.
But look into CAMM modules and LPCAMMS. My hope is that we will get modular LPDDR5X-8533 on AMD Strix Halo boards.
GDDR is actually super cheap! I think it would only be like another $75 on paper to double the 4090’s VRAM to 48GB (like they do for pro cards already).
Nvidia just doesn’t do it for market segmentation. AMD doesn’t do it for… honestly I have no idea why? They basically have no pro market to lose, the only explanation I can come up with is that their CEOs are colluding because they are cousins. And Intel doesn’t do it because they didn’t make a (consumer) GPU that was eally worth it until the B580.
The funny thing is James Cameron acts like the movie and message are so profound.
The issue with Macs is that Apple does price gouge for memory, your software stack is effectively limited to llama.cpp or MLX, and 70B class LLMs do start to chug, especially at high context.
Diffusion is kinda a different duck. It’s more compute heavy, yes, but the “generally accessible” software stack is also much less optimized for Macs than it is for transformers LLMs.
I view AMD Strix Halo as a solution to this, as its a big IGP with a wide memory bus like a Mac, but it can use the same CUDA software stacks that discrete GPUs use for that speed/feature advantage… albeit with some quirks. But I’m willing to put up with that if AMD doesn’t price gouge it.
second-hand TPU
From where? I keep a look out for used Gaudi/TPU setups, but they’re like impossible to find, and usually in huge full-server configs. I can’t find Xeon Max GPUs or CPUs either.
Also, Google’s software stack isn’t really accessible. TPUs are made for internal use at Google, not for resale.
You can find used AMD MI100s or MI210s, sometimes, but the go-to used server card is still the venerable Tesla P40.
Thats perfect.
Avatar the Last Airbender, and the Legend of Korra.
The animation, specifically. The Netflix series is OK, and there is absolutely no movie adaptation.
Funny enough, I watched them when I was like 19, but Korra, Zuko and Iroh are still three of my favorite characters in anything ever. I remember it vividly, where I can’t remember a single line from the James Cameron Avatar (which I watched as a teen when it came out).
James Cameron’s Avatar series.
Then again… Does anyone actually like it? It seems to have all this online hype when it’s such a boring visual spectacle.
It’s like the opposite of the other Avatar franchise, which wasn’t a commercial hit, and seems less popular on paper, but seems to have a massive cultural impact.
Hard disagree, lol. That’s a classic.
Yep. I didn’t mean to process shame you or anything, just trying to point out obscure but potentially useful projects most don’t know about :P